Search for: All records

Subduction zones are fundamental features of Earth's mantle convection and plate tectonics, but mantle flow and pressure around slabs are poorly understood because of the lack of direct observational constraints on subsurface flow. To characterize the linkages between slabs and mantle flow, we integrate high‐resolution representations of Earth's lithosphere and slabs into a suite of global mantle convection models to produce physically plausible present‐day flow fields for Earth's mantle. We find that subduction zones containing wide, thick, and long slabs dominate regional mantle flow in the neighboring regions and this flow conforms to patterns predicted by simpler regional subduction models. These subduction zones, such as Kuril‐Japan‐Izu‐Bonin‐Mariana, feature prismatic poloidal flow coupled to the downgoing slab that rotates toward toroidal slab‐parallel flow near the slab edge. However, other subduction zones, such as Sumatra, deviate from this pattern because of the competing influence of other slabs or longer‐wavelength mantle flow, showing that upper mantle flow can link separate subduction zones and how flow at subduction zones is influenced by broader scale mantle flow. We find that the non‐linear dislocation creep reduces the coupling between slab motion and asthenospheric flow and increases the occurrence of non‐ideal flow, in line with inferences derived from seismological constraints on mantle anisotropy.

The planform rearrangement of river basins is recognized as an important process for landscape evolution. The boundaries of river basins can shift either through gradual drainage divide migration or discrete river captures, but the methods for identifying these processes often rely on topographic evidence that remains otherwise untested. Moreover, efforts to understand the relative importance of either process are hampered by a lack of age constraints on river captures. We use¹⁰Be‐derived erosion rates to test whether, and how, divide motion is occurring at three locations along the Blue Ridge Escarpment in the Appalachian Mountains. In the Pee Dee River basin, we find that the escarpment is migrating inland up to 45 m/Myr, consistent with topographic evidence for gradual divide migration. In the Dan River basin, erosion rates support the topographic evidence for river capture, and we use a forward model of river incision to estimate that the capture likely occurred in the past 12.5 Myr. In the South Fork Roanoke River basin, where the presence of a knickzone has been interpreted as evidence that a river capture initiated a pulse of faster erosion, we instead measure nearly uniform tributary erosion rates above and within the mainstem knickzone. Simulations show that river incision into a more erodible layer of rock, with or without a river capture, could produce the observed topography and erosion rates in the South Fork Roanoke River. Our results show how multiple lines of evidence can illuminate the rates and mechanisms of river basin reorganization.